摘要
基于已有的相关PLS算法,提出了针对QSAR研究和工业过程控制建模的环境要求的PLS回归改进算法:加强递归PLS算法.模拟实验结果表明:在实时建模过程中,该算法的性能优于传统的PLS回归算法.
Based on the related PLS algorithms,a new improved recursive exponentially weighted PLS regressions algorithms was derived for the QSAR research and industrial process control modeling.Simulation experiments show that in the real-time modeling process,the performance of this algorithm is superior to the traditional PLS regression algorithm.
出处
《江西师范大学学报(自然科学版)》
CAS
北大核心
2012年第6期626-630,共5页
Journal of Jiangxi Normal University(Natural Science Edition)
基金
国家"973"计划(2010CB530602
2010CB530603)
国家"863"计划(2012AA02A609)
国家自然科学基金(81160424)资助项目
关键词
偏最小二乘法回归
kernel算法
算法改进
加权递归算法
partial least squares(PLS)regression
kernel algorithm
algorithms improvement
recursive exponentially weighted algorithms